JPH1169165A - Character image coding method and system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a character image coding method and its system where an input character image is coded to obtain a decoded image with high image quality. SOLUTION: A threshold level improper point detector 7 obtains a difference between a pixel and an adjacent pixel and selects a pixel value between the pixel and tone pixel value of the adjacent pixel with a small difference as a threshold level improper value and adds number of times where the pixel values are discriminated to be the threshold level improper value. A threshold level improper point detection section 7 applies the operation above to all pixels to detect the threshold level improper point information with respect to an input image. A threshold level revision section 8 compares the detected threshold level improper point information with a threshold level improper point of a pixel value close to the threshold level among the threshold level improper points decided with a threshold level decision section 4 within a prescribed range and a quantization threshold level is revised to the pixel value when the pixel value with a smaller threshold level improper point is in existence in the range.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character image encoding method, and more particularly, to a character image encoding method and apparatus for compressing and encoding a character image using multi-value run-length encoding.

[0002]

2. Description of the Related Art FIG. 5 is a block diagram showing an example of a conventional image coding apparatus. In FIG. 1, character image data such as subtitles of a movie, for example, input from an input terminal 1 is supplied to and stored in a frame memory 2 and then read out to a quantizer 3 and a threshold determination unit 4 respectively. Supplied. The threshold determining unit 4 determines a quantization threshold according to the input character image data, and supplies the quantization threshold to the quantizer 3.

[0003] Thus, character image data quantized to a plurality of gradations is taken out from the quantizer 3 and supplied to the run-length encoder 5, where it is determined according to the combination of the data element and the number of appearances. The data is encoded by a known run-length encoding method for variable-length encoding, compressed, and output to the output terminal 6 as a bit stream.

[0004]

However, in the above-described conventional image coding apparatus, since the character image data is quantized using a fixed threshold value, a portion where the gradation changes delicately is used. Pixel value and the quantization threshold may coincide with each other, and therefore, there is a problem that an unnatural level difference is likely to occur in the gradation in the decoded image.

[0005] The present invention has been made in view of the above points,
It is an object of the present invention to provide a character image encoding method and apparatus capable of encoding an input character image so as to obtain a high-quality decoded image.

[0006]

In order to achieve the above object, a character image encoding method according to the present invention determines a quantization threshold value of input character image data and determines a pixel value of a pixel of interest and its surrounding pixels. The difference value is compared in magnitude with a predetermined set value, the quantization threshold determined according to the comparison result is adaptively changed, and the input character image data is converted into a plurality of gradations based on the changed quantization threshold. , And then run-length encoded and compressed.

The character image coding apparatus of the present invention further comprises: a threshold value determining unit for determining a quantization threshold value of input character image data; and a difference value between a pixel of interest and a peripheral pixel thereof in the input image data. Threshold value changing means for comparing the difference value with a predetermined set value, and adaptively changing a quantization threshold value determined according to the comparison result; and input characters based on the quantization threshold value changed by the change means. It has a configuration including a quantizer for quantizing image data into a plurality of gradations, and an encoder for performing run-length encoding, compressing and outputting data output from the quantizer.

According to the method and apparatus of the present invention, a difference value between a pixel value of a pixel of interest and its surrounding pixels is obtained in input image data, the difference value is compared in magnitude with a predetermined set value, and a quantum value is determined according to the comparison result. Since the quantization threshold is adaptively changed, the quantization threshold can be hardly set to the pixel value of a portion where the difference value is slightly smaller than the set value and the gradation is slightly changed.

[0009]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a character image encoding apparatus according to the present invention. 5, the same components as those in FIG. 5 are denoted by the same reference numerals. The embodiment shown in FIG. 1 is an example in which input character image data is quantized to quaternary values and then run-length encoded. The conventional apparatus shown in FIG. The feature is that a change unit 8 is added.

Next, the operation of this embodiment will be described. Character image data such as subtitles of a movie, which is input via the input terminal 1, is supplied to and stored in the frame memory 2, and is read out to be quantized 3, the threshold determination unit 4, and the threshold improper. Each is supplied to the point detection unit 7. The threshold determination unit 4 determines a quantization threshold according to the input character image data, and supplies the quantization threshold to the threshold changer 8 instead of the quantizer 3. Here, the quantization threshold value determined by the threshold value determination unit 4 is such that a point dividing the range between the maximum value and the minimum value of the pixel values of the input image into four equal to the threshold value a, the threshold value b, and the threshold value c.

The output character image data of the frame memory 2 is also supplied to the threshold improper point detection unit 7 at the same time. The threshold improper point detector 7 calculates the coordinates (x, y)
And the coordinates (x, y−1), (x, y + 1), (x−1,
y) and a difference value between four adjacent pixels (x + 1, y) are determined, and a pixel value between the pixel value of the adjacent pixel having a small difference value and the pixel of interest is set as a threshold ineligible value, and The number of times that the value is determined to be the threshold ineligible value is added. here,
Coordinates (x, y-1), (x, y + 1), (x-1, y)
When the difference value between the pixel of interest and the pixel of interest (x + 1, y) and the pixel of interest is equal to or less than “10”, the pixel value between them is detected as a threshold-eligible pixel. Also, the number of times a pixel is determined to be a threshold inappropriate pixel is referred to as a threshold inappropriate point of the pixel value.

Next, a pixel of interest at coordinates (x, y) and coordinates (x, y-
1), (x, y + 1), (x-1, y) and (x + 1,
y), and four adjacent pixels (x, y−2), (x, y + 2), (x−2, y),
A difference value from (x + 2, y) is observed, and the value is “10”.
In the following cases, the threshold inappropriate point is added to the pixel value between the two pixel values. Further, observation is performed by extending the range in the vertical and horizontal directions until a pixel having a difference value of “10” or more appears. Threshold improper point detection unit 7
Performs the above operation on all pixels, and detects threshold improper point information on the input image.

The detected threshold improper point information is given to the threshold changing unit 8. The threshold value changing unit 8 includes a threshold value inappropriate point of the threshold value a determined by the threshold value determining unit 4 and a threshold value a.
Are compared within a predetermined range. Here, it is assumed that the comparison is performed in a range of 10 levels from the threshold value a in each of the plus direction and the minus direction. As a result of the comparison, when a pixel value having a threshold inappropriate point smaller than the threshold inappropriate point of the threshold a exists in the range, the threshold a is changed to the pixel value and set as the threshold a ′. This is also performed for the thresholds b and c. The changed threshold value a ′, threshold value b ′, threshold value c ′ and the input character image data are input to the quantizer 3.

The quantizer 3 codes pixels below the threshold value a 'with a code "00", pixels larger than the threshold value a' and below the threshold value b 'with a code "01", larger than the threshold value b', and
Pixels equal to or smaller than the threshold value c 'are represented by a code "10", and pixels larger than the threshold value c' are represented by a code "11". The quaternary quantized character image data is supplied to a run-length encoder 5, where it is run-length encoded and then output to an output terminal 6.

Next, an example of run-length encoding will be described with reference to FIG. When one to three consecutive values are the same, the number of consecutive pixels is represented by 2 bits as indicated by reference numerals 21 and 22 in FIG.
Expressed in bits. Next, when the same value continues from four to fifteen, as shown at 23 in FIG.
After arranging 0 ", a 4-bit area 24 indicating the number of consecutive pixels and a 2-bit area 25 indicating the pixel value are sequentially arranged.

When the same value continues from 16 to 63, as shown in FIG. 2C, a 6-bit area 26 indicating the number of consecutive pixels follows the 4-bit "0" area 26. An area 27 and a 2-bit area 28 indicating a pixel value are sequentially arranged. Similarly, when the same value continues from 64 to 255, as shown in FIG.
Following the area 29, an 8-bit area 30 indicating the number of consecutive pixels and a 2-bit area 31 indicating the pixel value are sequentially arranged. Similarly, when the same value continues from 256 pixels to the end of the pixel line, as shown in FIG. 2E, a 14-bit "0" area 32 is followed by 2 indicating the value of the pixel. The bit area 33 is arranged.

Next, the above embodiment will be described using a specific example. In FIG. 1, the maximum pixel value of the character image data input to the threshold determination unit 4 is “255”, and the minimum value is “0”. The threshold value determining unit 4 divides the pixel value between the maximum value and the minimum value into four equal values, that is, “64”, “128”
And “192” are determined as a threshold a, a threshold b, and a threshold c, respectively. At the same time, the threshold improper point detection unit 7 observes the difference value between the target pixel and its adjacent pixel in the input character image data. Here, the difference value between the target pixel and its adjacent pixels in the vertical and horizontal directions is observed, and a threshold inappropriate point for the target pixel is detected. In addition, the range in which the difference value is observed is set until no inappropriate threshold pixel is detected.

FIG. 3 shows the pixel of interest at the coordinates (x, y) and the pixel values of its surrounding pixels. As shown in the figure, the pixel value of the target pixel is “194”, and the coordinates (x, y−
Since the pixel value of the adjacent pixel of 1) is “235”, their difference value is “41” and the coordinates (x, y +
Since the pixel value of the adjacent pixel of 1) is “18”, the difference value from the pixel of interest is “176”, and these are all “1”.
Since it is larger than “0”, the threshold inappropriate pixel value is not detected.

On the other hand, since the pixel value of the adjacent pixel at the coordinates (x-1, y) is "192", the difference value from the pixel of interest is "2", and the pixel value between them is "192". ,
The threshold inappropriate point is added to “192” and “193” with “193” as the threshold inappropriate pixel value. Since the pixel value of the adjacent pixel at the coordinates (x-1, y) is "192" and the pixel value of the adjacent pixel at the coordinates (x-2, y) is "188", their difference value is "4". , The threshold value inappropriate points of the pixel values “188”, “189”, “190”, and “191” are similarly added. Further, the coordinates (x−3,
Since the pixel value of the pixel y) is “151”, the coordinates (x−
The difference value between the adjacent pixel of (2, y) is “37” and “1”
0 ”, the threshold inappropriate pixel value is not detected.
Therefore, the observation of the difference value of the pixel of interest in the left direction ends here.

Next, since the difference value between the pixel of interest and the adjacent pixel at the coordinate (x + 1, y) whose pixel value is "187" is "7", the pixel values "187", "188",
The threshold inappropriate points of “189”, “190”, “191”, “192”, and “193” are added. Further, since the difference value between the adjacent pixel at the coordinate (x + 1, y) and the pixel at the coordinate (x + 2, y) whose pixel value is “172” is “15”, the threshold inappropriate pixel is not detected. Table 1 shows the threshold inappropriate points in the case of FIG. 3 obtained as described above.

[0022]

[Table 1] This operation is performed for all pixels, and a threshold inappropriate point is detected for each pixel.

The threshold changing unit 8 changes the threshold in response to the threshold improper point detected by the threshold improper point detecting unit 7. Here, the threshold c at the pixel of interest
Will be described as an example. In addition, the range for comparing the threshold improper point is to be compared in the range of 5 levels in both the plus direction and the minus direction from the threshold value c. That is,
Since the threshold value c is “192” as described above, the threshold values in the range from the pixel values “187” to “197” are compared, and the pixel value having the smallest threshold inappropriate point is determined as the threshold value c ′.
And

From Table 1, since the threshold value inappropriate point is the lowest in this range and the pixel value close to the threshold value c is "194", the threshold value c of the target pixel at the coordinates (x, y) is set to "192". To “194”. This is called threshold c
(X, y). Similarly, a threshold value a (x, y) and a threshold value b (x, y) are obtained.

Next, the quantizer 3 calculates these quantization thresholds a
Using (x, y), b (x, y) and c (x, y),
The character image data read from the frame memory 2 is quantized. Table 2 shows the quantization table.

[0026]

[Table 2] The pixel value of the target pixel is “194” and the threshold value c
Since (x, y) is “194”, it is represented by the code “10” as can be seen from Table 2. The coordinates (x−
The pixel value of the adjacent pixel on the left of the target pixel of (1, y) is “19”.
2 ", the pixel value of the adjacent pixel on the right side of the pixel of interest at the coordinates (x + 1, y) is" 187 ", and the quantizer 3 outputs the code" 10 "with reference to the quantization table of Table 2. I do. The quantizer 3 outputs the code of the value shown in FIG. 4 by referring to the quantization table for other pixels in the same manner.

Conventionally, if the quantization threshold value is not changed by the threshold improper point, the threshold value c is "192". Therefore, the target pixel is quantized to code "11", and the left and right adjacent pixels are quantized. The values after the conversion are different, and an unnatural level difference occurs in the decoded image. On the other hand, in this embodiment, as can be seen from FIG. 4, pixels adjacent to the left and right of the target pixel at the coordinates (x, y) have the same code “1” as the target pixel.
Since it is quantized to 0 ", an unnatural level difference does not occur in the decoded character image.

As described above, in this embodiment, the threshold value inappropriate point is detected, and the quantization threshold is changed according to the threshold value inappropriate point. Is hardly set, and an unnatural level difference does not occur in the gradation of the decoded image.

[0029]

As described above, according to the present invention,
According to the method and apparatus of the present invention, a difference value between a pixel value of a target pixel and its surrounding pixels in input image data is obtained, the difference value is compared with a predetermined setting value, and a quantization threshold is set according to the comparison result. By adaptively changing, it is difficult to set a quantization threshold value for a pixel value of a portion where a difference value is slightly smaller than a set value and a gradation is slightly changed, so that an unnatural level difference is hardly generated in a decoded image gradation. Thus, a high-quality decoded character image can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is an explanatory diagram of an example of run-length encoding.

FIG. 3 is a diagram illustrating an example of a pixel value of a character image.

FIG. 4 is a diagram illustrating an example of quantized data of the character image in FIG. 3;

FIG. 5 is a block diagram of an example of the related art.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Character image data input terminal 2 Frame memory 3 Quantizer 4 Threshold decision unit 5 Run-length encoder 6 Encoded data output terminal 7 Threshold improper point detection unit (threshold changing means) 8 Threshold changing unit (threshold changing means)

Claims (5)

[Claims] 1. A quantization threshold value of input character image data is determined, a difference value between a pixel value of a target pixel and a pixel value of a peripheral pixel is compared with a predetermined setting value, and the determined value is determined according to a comparison result. A character image characterized by adaptively changing a quantization threshold, quantizing the input character image data into a plurality of gradations based on the changed quantization threshold, and then performing run-length encoding and compression. Encoding method. 2. A method for determining a quantization threshold value of input character image data, comparing a difference value between a pixel value of a target pixel and a pixel value of a peripheral pixel thereof with a predetermined set value, and determining that the difference value is larger than the set value. When the pixel value is small, the pixel value between the compared two pixel values is regarded as an ineligible value, the number of times that the pixel value is determined as an ineligible value is added, and the number of ineligible value judgments on the entire screen is detected. The quantization threshold is changed to a value corresponding to a pixel value having the smallest number of times of inappropriate value detection detected within a preset range, and a plurality of the input character image data are set based on the changed quantization threshold. A character image encoding method characterized in that the image is quantized to the grayscale of, and then run-length encoded and compressed. 3. A threshold value determining unit for determining a quantization threshold value of input character image data, a difference value between a pixel value of a target pixel and a pixel value of a peripheral pixel in the input image data, and the difference value is set to a predetermined value. Threshold changing means for adaptively changing the determined quantization threshold in accordance with the comparison result; and a plurality of floors for the input character image data based on the quantization threshold changed by the changing means. A character image encoding device, comprising: a quantizer for quantizing in a key; and an encoder for performing run-length encoding, compressing and outputting data output from the quantizer. 4. The threshold value changing means calculates a difference value between a pixel value of a target pixel and a pixel value of a peripheral pixel in the input image data, compares the difference value with a predetermined set value, and compares the difference value with a predetermined set value. When the difference value is small, it is compared by 2
A pixel value between two pixel values is regarded as an unqualified value, the number of times that the pixel value is determined as an unqualified value is added, and a detection unit that detects the number of unqualified value determinations on the entire screen, and the threshold value determining unit And a threshold changing unit that changes the determined quantization threshold to a value corresponding to a pixel value having the least number of times of the inappropriate value determination detected within a range set in advance by the detection unit. Item 3. The character image encoding device according to Item 3. 5. The threshold determining unit, wherein the detecting unit sets a range of peripheral pixels for which the difference value is obtained between the target pixels to peripheral pixels from which a pixel value determined to be an inappropriate value is not detected. When quantizing to a 2 ^m (m is a natural number) value, (m−1) pixel values that divide the range between the maximum pixel value and the minimum pixel value into m equal parts are determined as quantization threshold values. The character image encoding device according to claim 4, wherein